confused about bit strings
Is the following behavior intentional? If so, what's the reasoning
behind it?
nconway=# select 1::bit;
bit
-----
0
(1 row)
nconway=# select '1'::bit;
bit
-----
1
(1 row)
nconway=# select X'1'::bit;
bit
-----
0
(1 row)
nconway=# select 1::bit varying;
ERROR: cannot cast type integer to bit varying
nconway=# select 4::int2::bit;
ERROR: cannot cast type smallint to bit
nconway=# select 4::bit;
bit
-----
0
(1 row)
nconway=# select '4'::bit;
ERROR: "4" is not a valid binary digit
nconway=# select X'4'::bit varying;
varbit
--------
0100
(1 row)
-- why is that 4 bits, not 3?
nconway=# select '14'::int::bit;
bit
-----
0
(1 row)
nconway=# select bit('14'::int);
ERROR: syntax error at or near "'14'" at character 12
nconway=# select "bit"('14'::int);
bit
----------------------------------
00000000000000000000000000001110
(1 row)
-- shouldn't bit be equivalent to bit(1), which should be
right-truncated?
Cheers,
Neil
Neil Conway writes:
nconway=# select 1::bit;
bit
-----
0
(1 row)
Oops. I've always thought that casting between int and bit should be
disallowed, but apparently it keeps sneaking back in.
nconway=# select X'4'::bit varying;
varbit
--------
0100
(1 row)
-- why is that 4 bits, not 3?
SQL says so:
12) The declared type of a <hex string literal> is fixed-length
bit string. Each <hexit> appearing in the literal is equivalent
to a quartet of bits: 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, A, B, C,
D, E, and F are interpreted as 0000, 0001, 0010, 0011, 0100,
0101, 0110, 0111, 1000, 1001, 1010, 1011, 1100, 1101, 1110,
and 1111, respectively. The <hexit>s a, b, c, d, e, and f have
respectively the same values as the <hexit>s A, B, C, D, E, and
F.
nconway=# select "bit"('14'::int);
bit
----------------------------------
00000000000000000000000000001110
(1 row)
-- shouldn't bit be equivalent to bit(1), which should be
right-truncated?
It is, but here you're calling a function, not referring to the type.
--
Peter Eisentraut peter_e@gmx.net